Ryan Goulden

So, this change could be refactored to not add `ThawAll()` as a method everywhere, and instead use the normal `Set` mechanism with a new `configs.ThawedRecursive` enum. I think this would...

~It looks like the new cgroups objects have had the corresponding changes made for them as well here. (Not totally sure since I'm unfamiliar with the new cgroups stuff.) @mruck...

> My worry with this change is that people can setup sub-cgroups that maybe related to the container but not bound to its lifecycle and this would be a breaking...

Yes. (Although, in our case, we're not intentionally suspending containers. The temporary freezing and unfreezing that runC does while sending signals is enough to cause the problem to exhibit from...

Rebases have not been entirely trivial since underlying cgroupsv2 code keeps changing, so it's probably not worth the effort until there's enough interest in reviewing the bugfix anyways ¯\\_(ツ)_/¯

Ah, cool. Yeah, if things actually start moving here I'm happy to deal with the conflicts.

I think the assert is correct. I believe the code is actually accessing `a + 1` (aka only words before `a + 2`). We invoke _nn_divapprox_helper with `a = a-s`...

:+1: (FYI: dunno if I mentioned it, this and the other issue I filed came from me compiling & running tests with some -fsanitize args. Mostly because I was trying...

sub_80480d0 of [KPRCA_00009](https://github.com/Vector35/binaryninja-api/files/241899/KPRCA_00009.zip) basically exhibits every CFG-routing oddity I've ever seen. Some edges with (visually) distinct issues may be found at 0x08048119, 0x0804821d, 0x080480ed, 0x080481f1 Particularly fun edge at 0x0804836d,...

@nedwill Sure, I can take a look. In case it matters, can you make a note of your OS X version and `clang --version` (or whatever it was that you...